Skip to content
@Bayanat-Labs

Bayanat Labs

The leading Arabic data company for AI, based in Riyadh. Open tools, benchmarks and evaluation kits for Arabic and its dialects.
Bayanat Labs logo

Bayanat Labs

بيانات لابز

Arabic AI shouldn't sound translated.

Website · Run a pilot · LinkedIn · hello@bayanatlabs.com


Bayanat Labs is a leading Arabic data company for AI, based in Riyadh. We work only on data, from sourcing and annotation to alignment (RLHF and human feedback) and evaluation. We are model-agnostic and never compete with our clients' models. Our data covers 25+ Arabic varieties and dialects and is produced by vetted native speakers, linguists and licensed domain professionals.

What you will find here

This is where we publish open code for Arabic AI: benchmark harnesses, evaluation kits, validators and the scripts that rebuild our open datasets. Each project repository ships with a license and a CITATION.cff file, and links to its page on bayanatlabs.com.

Large and third-party datasets do not live in git. Small fixtures we wrote ourselves, such as the Arabic Retrieval Starter Kit, can. Each repository says where its data comes from.

Featured work

  • LabelBench · Repository
    A reproducible audit of automated Arabic data labeling. On the fixed held-out manifest, a local language model reaches 78% exact-match accuracy on binary Arabic toxicity. On the full seven-label taxonomy it falls to 25.3%, and on five dialect regions it reaches 22.4%, where chance is 20%. Read the paper (PDF).

  • Arabic Retrieval Starter Kit · Repository
    A small, fully synthetic Arabic and English RAG evaluation kit. Gulf and MSA questions, graded relevance judgments, outdated-policy hard negatives and unanswerable questions. Free under CC BY 4.0.

  • Arabic Agent Reliability Lab · Repository
    A methods beta for state-based evaluation of Arabic AI agents, with evidence receipts. A fluent answer is not a completed transaction.

Work with us

Contribute and cite

Bug reports and requests for open Arabic datasets, tools and benchmarks are welcome as issues on the relevant repository. Please read the contributing guide first. To cite our work, use the "Cite this repository" button on any project repository.

Contact

Email hello@bayanatlabs.com, use the contact form on our website, or follow Bayanat Labs on LinkedIn.


بيانات لابز

بيانات لابز (Bayanat Labs) شركة رائدة في البيانات العربية للذكاء الاصطناعي، مقرّها الرياض، تعمل في جمع البيانات وتوسيمها ومواءمتها وتقييمها. نعمل على البيانات فقط، ونحن محايدون تجاه النماذج ولا ننافس نماذج عملائنا. ينفّذ العمل متحدثون أصليون ولغويون ومختصون مرخّصون بعد اجتياز الفرز، عبر أكثر من 25 لهجة وتنويعة عربية.

ننشر هنا شيفرة مفتوحة المصدر للذكاء الاصطناعي العربي، منها أدوات تشغيل المعايير (benchmark harnesses) وحزم التقييم وأدوات التحقق من صحة البيانات، والنصوص البرمجية التي تعيد بناء مجموعات بياناتنا المفتوحة. ولكل مستودع من مستودعات مشاريعنا رخصة وملف CITATION.cff ورابط إلى صفحته على bayanatlabs.com.

أما مجموعات البيانات الكبيرة أو المملوكة لجهات أخرى فلا نضعها في git، وقد نضع فيه مجموعات صغيرة كتبناها بأنفسنا، مثل Arabic Retrieval Starter Kit. ويذكر كل مستودع مصدر بياناته.

للعمل معنا، ابدأ بمشروع تجريبي محدد النطاق على بياناتك، يُنجَز عادةً خلال أسبوعين، ويشمل ضبط الجودة على مجموعة ذهبية وتقرير جودة.

اطلب مشروعًا تجريبيًا · المنتجات · المقالات · شبكة المواهب · LinkedIn · hello@bayanatlabs.com

Pinned Loading

  1. arabic-retrieval-starter-kit arabic-retrieval-starter-kit Public

    A small, synthetic Arabic and English RAG evaluation kit from Bayanat Labs, with 12 passages, 12 queries, graded qrels, outdated-policy hard negatives and unanswerable questions. CC BY 4.0.

  2. arabic-agent-reliability-lab arabic-agent-reliability-lab Public

    Arabic Agent Reliability Lab by Bayanat Labs, a local harness for evaluating Arabic tool-using AI agents in a synthetic commerce sandbox. It grades tool calls and the final state, reports communica…

    Python

  3. labelbench labelbench Public

    LabelBench by Bayanat Labs, a reproducible audit of automated Arabic data labeling. Measures how much of a labeling queue can be accepted at a declared quality floor, by taxonomy, task and Arabic d…

    Python

Repositories

Showing 10 of 17 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…